REVIEW OF HIPPOCAMPUS — GINSBURG 583 



Trunk deep ; snout of medium length. Filaments short, more or less 

 branched, present on spines of head and of upper ridge of trunk and 

 anterior part of tail. Brown, with lighter blotches around bases of 

 spines of trunk, the blotches coalescent (the color pattern somewhat 

 as in specimens of hudsonius or pundulatus of similar size) ; white dots 

 present, but scanty; dorsal with obliquely longitudinal rows of rather 

 faint brownish spots near base, no submarginal band. 



Measurements. — Length 68 mm, without any trace of a brood 

 pouch; depth 17, snout 8.5, postorbital 12, head 24, trunk 38, tail 56.5, 

 and orbit 4.5 percent of length. 



Distinctive characters and relationships. — The foregoing account is 

 based on a single specimen that I refer with considerable doubt to 

 Giinther's species, which is also known from but one specimen. The 

 species of Hippocampus are so variable intraspecifically, and so closely 

 approaching or even overlapping interspecifically, that it seems fool- 

 hardy to base a species on a single specimen, except where it shows 

 some salient character unmistakably distinguishing it. There must 

 be even greater uncertainty to attempt to identify a single specimen 

 with a poorly established species without comparing it directly with 

 the type. However, this specimen is evidently of a different species 

 from any of the others from the American coasts described in the 

 present paper, and it agrees fairly well with the inadequate account 

 of villosus, except that Giinther's specimen apparently had a longer 

 snout. Not wishing to establish a new species on a single specimen in 

 this case, I provisionally refer it to villosus. 



Judged from the species from the American coasts known at present, 

 this specimen belongs to a species nearest to reidi on one hand and to 

 punctulatus on the other, but it apparently differs from both. The 

 most striking feature is its relatively small number of segments, both 

 trunk and caudal segments. The 10 trunk segments represent the 

 most usual number found in the subgenus Jamsus. Of the other 

 species described herein, only one specimen of hudsonius, out of 76 

 examined, had this number, while in all the rest of the species not 

 one specimen was found with 10 trunk segments. It is possible that 

 the specimen here referred to villosus represents a rare variant, but 

 the probabilities are much more strongly in favor of its representing 

 a species that normally has fewer trunk segments. The number of 

 caudal segments is also near to the normal condition in the subgenus 

 Jamsus, but it also falls at the extreme of the frequency distributions 

 of reidi and punctulatus (compare with table 1 ) . This specimen further 

 differs from reidi in its deeper body and strilvingly better developed 

 tubercles, and from punctulatus in having a deeper body when speci- 

 mens of approximately the same size are compared (see table 3). 

 From the two species belonging to the subgenus Jamsus it differs 



